Voice recognition in Sync with simplicity

Following is one of three recently posted stories from the special digital edition on Automotive Intelligence. The other two are on Cleantech power requirements and Automotive MEMS.-- Nicolas Mokhoff, ed.


Ford Motor Co. claims to have tamed voice recognition with the latest version of its Sync technol­ogy by upping the number of com­mands a hundredfold while simplifying their grammar.

Sync technology is used to connect mobile devices, such as smartphones, to an in-­vehicle network that is con­trolled with voice commands. The im­proved Sync now allows any command to be directly spoken and recognized in natural language, whereas with most other voice recognition systems, a speaker must identify the pertinent menu of commands before issuing the desired command.

"We have flattened the hierarchy of the grammar," said electrical engineer Jim Buczkowski, a Henry Ford Technical Fellow and director of global electrical and electronics systems engineering for Ford. “Now there are many more commands at the first level. For instance, instead of saying 'phone' first, then saying who you want to call, now you can just say, 'Call Jim.'"

Ford Sync was co­-developed with Microsoft on the Windows Embedded Automotive software platform and uses voice recognition algorithms licensed from Nuance Communications Inc. The Sync technology was recently integrated with MyFord Touch, which uses a dash-­mounted resistive touchscreen to control vehicle functions.

"We work with many automobile makers, but the Ford Sync system is probably our most sophisticated implementation to date," said Brian Radloff, director of automotive solutions at Nuance. "For instance, on the destination entry side you can now speak an entire address in a single utterance– such as 100 Main St., Detroit, Michigan– instead of having to speak the street number, street name, city and state separately.

"Sync also uses our lat­est algorithms for on­-the-­fly adaptation, so its accuracy increases the more you use it."

Most functions—from tuning the radio to asking for directions or even call­ing 911—can now be con­trolled by touch or voice. The second­-generation Sync also recognizes thou­sands of aliases that make access more convenient for drivers by not requiring them to remember specific commands. For instance, saying "warmer," "increase temperature" or "temperature up" will all bump up the thermometer reading in the cabin. Simplifications to Sync also help drivers control their own mo­bile devices with voice.

For instance, when you first buy a Sync­-equipped vehicle it will automatically index the names and addresses on your smartphone as well as the names of the bands and indi­vidual tracks on your music player. Within a few minutes, those indexed entries can be accessed by voice commands.
Further, the system will con­vert text messages to audio and read them using Nuance's text­-to­-speech engine. Drivers can "text back" via voice con­trol by choosing one of up to 20 predefined text responses.

Ford has also been busy persuading developers of such popular Web applications as MapQuest and Google Maps to add "Send to Sync" buttons to their Web sites. Once you find a destination using the map­ping program on your PC, all you need to do is hit the "Send to Sync" button and fill in your cell phone number. Then, when you enter your Ford vehicle, the map you selected in MapQuest or Goggle Maps will automati­cally be uploaded to the dash­board display.

Ford’s "Traffic, Directions and Information" integrates its voice command structure with GPS and Bluetooth connectivity to provide personalized traffic reports; turn­by­turn directions; and instant access to news, sports and weather. Destination-related queries can be made, such as "Find a hotel."
Ford provides in­-vehicle Wi­Fi service for any driver who plugs a 2G, 3G or 4G modem into a special USB slot. Unlike GM’s OnStar system, which in­cludes a cell phone connec­tion on the vehicle, Ford's Sync depends on the connec­tivity link provided by the user’s cell phone.

Ford is also allowing third­-party developers to target its in-­vehicle network with the AppLink application program­ming interface (API), which lets apps add unique capabili­ties that can be controlled via voice commands.

"For instance, Pandora’s app has added our API so that you can use Sync to stream music from the cloud in your car with-voice commands," said Buczkowski. Several other apps are already Sync compatible, such as the Stitcher app for radio news and the OpenBeak app for Twitter, with more slated to add AppLink compatibility this year.

The voice recognition technology is currently fluent in English, French and Spanish, and Ford plans to expand Sync’s vocabulary to 21 languages next year. Sync has been installed in more than 3 million Ford vehicles and will be made available on 80 percent of the company’s production vehicles over  the next four years, according to Buczkowski.

More from this article and other articles can be found in the Special digital edition on Automotive Intelligence.